Inference of Naturally Acquired Immunity Using a Self-matched Negative-Control Design

Area of Science:

• Epidemiology
• Biostatistics
• Immunology

Background:

• Host adaptive immunity is crucial for protection against recurrent infections.
• Estimating protective immunity often uses hazard ratios, but frailty bias due to individual susceptibility can confound results.
• Frailty bias arises from unmeasured individual heterogeneity, leading to spurious associations between prior infection and recurrence risk.

Purpose of the Study:

• To introduce a novel self-matched case-only inference method to control for unmeasured individual heterogeneity and frailty bias.
• To provide a flexible statistical framework for analyzing naturally acquired immunity and other exposures influencing recurrent event times.
• To derive a Mantel-Haenszel odds ratio for estimating the effect of prior infection on time to recurrence.

Main Methods:

• A self-matched case-only inference method is proposed, utilizing negative-control endpoints.
• The method compares event times for pathogen-specific endpoints and negative-control endpoints within counterfactual risk periods defined by infection history.
• A standard Mantel-Haenszel (matched) odds ratio is derived to quantify the effect of prior infection on recurrence.

Main Results:

• The proposed method effectively controls for unmeasured individual heterogeneity and frailty bias.
• Performance is compared against traditional proportional hazards modeling frameworks.
• The approach was applied to reestimate naturally acquired protection against rotavirus gastroenteritis, demonstrating its utility.

Conclusions:

• The self-matched negative-control design offers a robust alternative for analyzing naturally acquired immunity.
• This method can be applied to various exposures that influence the distribution of recurrent event times.
• Accurate estimation of protective immunity is essential for public health interventions and vaccine development.